检索结果-内蒙古大学图书馆

31st International Conference on Computational Linguistics, COLING 2025

作者： Zhang, Yuqi Ding, Liang Zhang, Lefei Tao, Dacheng School of Computer Science Wuhan University China The University of Sydney Australia College of Computing and Data Science Nanyang Technological University Singapore639798 Singapore

ISBN: (纸本)9798891761964

Aligning large language models (LLMs) with human values, particularly when facing complex and stealthy jailbreak attacks, presents a formidable challenge. Unfortunately, existing methods often overlook this intrinsic nature of jailbreaks, which limits their effectiveness in such complex scenarios. In this study, we present a simple yet highly effective defense strategy, i.e., Intention Analysis (IA). IA works by triggering LLMs' inherent self-correct and improve ability through a two-stage process: 1) analyzing the essential intention of the user input, and 2) providing final policy-aligned responses based on the first round conversation. Notably, IA is an inference-only method, thus could enhance LLM safety without compromising their helpfulness. Extensive experiments on varying jailbreak benchmarks across a wide range of LLMs show that IA could consistently and significantly reduce the harmfulness in responses (averagely -48.2% attack success rate). Encouragingly, with our IA, Vicuna-7B even outperforms GPT-3.5 regarding attack success rate. We empirically demonstrate that, to some extent, IA is robust to errors in generated intentions. Further analyses reveal the underlying principle of IA: suppressing LLM's tendency to follow jailbreak prompts, thereby enhancing safety. © 2025 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Counseling Responses for Mental Health Forum Questions with Early Maladaptive Schema Prediction 27

Counseling Responses for Mental Health Forum Questions with ...

引用

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Gollapalli, Sujatha Das Ang, Beng Heng Du, Mingzhe Ng, See-Kiong Institute of Data Science National University of Singapore Singapore Integrative Sciences and Engineering Programme National University of Singapore Singapore College of Computing and Data Science Nanyang Technological University China

ISBN: (纸本)9781643685489

State-of-the-art Large Language Models (LLMs) have shown remarkable capabilities for general Question Answering (QA) tasks. However, their practical use for answering mental health questions has been limited due to the missing link between LLM-generated answer responses and well-established theories and guiding principles from Psychology and Counseling. We present a first step in this direction with STeer, an AI-based method that supports Schema Therapy-enabled responses for mental health questions on community QA forums. STeer uses Early Maladaptive Schemas (EMSs), a fundamental concept from Schema Therapy that characterizes "self-defeating, unhealthy patterns of thought and behavior" in individuals, to effectively prompt state-of-the-art LLMs to generate specific, theoretically-grounded, empathetic counseling responses to mental health questions. We present EMSRank, a novel method based on the Personalized PageRank algorithm, to automatically predict the EMSs from mental health forum question texts. We show that EMSRank is computationally scalable and can be further combined with textual entailment to obtain high precision, explainable EMS labels for mental health forum questions. To address the current lack of annotated datasets, we also leveraged on EMSRank to create a first-of-its-kind, large dataset of about 23K EMS-annotated mental health questions from three diverse, currently operating, peer-support community forums for mental health. With the global rise in mental health issues, our work is a timely step towards enabling the use of AI-based assistive tools for counseling support on mental health community forums. © 2024 The Authors.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

Adversarial Machine Learning for Social Good: Reframing the Adversary as an Ally

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第9期5卷 4322-4343页

作者： Al-Maliki, Shawqi Qayyum, Adnan Ali, Hassan Abdallah, Mohamed Qadir, Junaid Hoang, Dinh Thai Niyato, Dusit Al-Fuqaha, Ala Division College of Science and Engineering Doha34110 Qatar Information Technology University Lahore54000 Pakistan University of New South Wales Sydney Sydney2052 Australia Qatar University Department of Computer Science and Engineering College of Engineering Doha34110 Qatar University of Technology Sydney School of Electrical and Data Engineering Sydney2052 Australia Nanyang Technological University College of Computing and Data Science 639798 Singapore

Deep neural networks (DNNs) have been the driving force behind many of the recent advances in machine learning. However, research has shown that DNNs are vulnerable to adversarial examples - input samples that have been perturbed to force DNN-based models to make errors. As a result, adversarial machine learning (AdvML) has gained a lot of attention, and researchers have investigated these vulnerabilities in various settings and modalities. In addition, DNNs have also been found to incorporate embedded bias and often produce unexplainable predictions, which can result in antisocial AI applications. The emergence of new AI technologies that leverage large language models (LLMs), such as ChatGPT and GPT-4, increases the risk of producing antisocial applications at scale. AdvML for social good (AdvML4G) is an emerging field that repurposes the AdvML bug to invent prosocial applications. Regulators, practitioners, and researchers should collaborate to encourage the development of prosocial applications and hinder the development of antisocial ones. In this work, we provide the first comprehensive review of the emerging field of AdvML4G. This paper encompasses a taxonomy that highlights the emergence of AdvML4G, a discussion of the differences and similarities between AdvML4G and AdvML, a taxonomy covering social good-related concepts and aspects, an exploration of the motivations behind the emergence of AdvML4G at the intersection of ML4G and AdvML, and an extensive summary of the works that utilize AdvML4G as an auxiliary tool for innovating prosocial applications. Finally, we elaborate upon various challenges and open research issues that require significant attention from the research community. © 2020 IEEE.

关键词： Taxonomies

来源：评论

学校读者我要写书评

暂无评论

INVESTIGATING THE IMPACT OF TREES AS NOISE BARRIER ON URBAN NOISE POLLUTION THROUGH 3D SCENE MODELLING 9

INVESTIGATING THE IMPACT OF TREES AS NOISE BARRIER ON URBAN ...

引用

9th International Conference on Big data Analytics, data Mining and Computational Intelligence, BigDaCI 2024, the 10th International Conference on Connected Smart Cities, CSC 2024 and the 16th International Conference on e-Health, EH 2024, Part of the 18th Multi Conference on Computer science and Information Systems 2024, MCCSIS 2024

作者： Yean, Seanglidet Heng, Zheng Ping Leow, Guan Wei Lee, Bu Sung Edwards, Peter College of Computing and Data Science Nanyang Technological University Singapore Manaaki Whenua – Landcare Research Wellington New Zealand

ISBN: (纸本)9789898704597

Noise pollution is a cause for concern, especially since it has detrimental health effects. Due to Singapore’s limited geographical size and rising population density, gas stations, roads, and even expressways are built close to residential neighborhoods. This situation is increasingly common in dense cities worldwide, making Singapore a relevant case study for global urban environments. Noise pollution in such areas is a significant problem, as a result of frequent traffic on these routes, especially during rush hours. The purpose of this project is to create a pipeline from a scene simulation to modelling noise propagation. This project aims to develop a scene modelling platform that will help bridge the gap between scene and noise modelling. It also aims to investigate different noise barriers and their effectiveness against vehicular noise pollution. Through experiments conducted in residential and park settings, we demonstrate the varying effects of trees on noise reduction, highlighting their potential to attenuate noise levels, especially in densely populated areas. Our findings reveal that while trees may not completely block noise like solid barriers, they offer additional benefits such as airflow and visual enhancement of the environment. Furthermore, we assess the impact of reforestation and strategic tree placement in reducing traffic noise, providing insights for urban planners and policymakers. Additionally, we present the case study modelling the scene after observed StreetView, showcasing the significant noise reduction benefits of planting trees in residential areas. Overall, our study contributes to the development of more effective strategies for mitigating noise pollution and creating healthier, more sustainable urban environments. Demo Video: https://***/zi1R-1xDkwU. © 2024 Big data Analytics, data Mining and Computational Intelligence. All Rights Reserved.

关键词： Noise abatement

来源：评论

学校读者我要写书评

暂无评论

Are LLMs Good Zero-Shot Fallacy Classifiers?

Are LLMs Good Zero-Shot Fallacy Classifiers?

引用

2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024

作者： Pan, Fengjun Wu, Xiaobao Li, Zongrui Luu, Anh Tuan Interdisciplinary Graduate Programme Nanyang Technological University Singapore College of Computing and Data Science Nanyang Technological University Singapore

ISBN: (纸本)9798891761643

Fallacies are defective arguments with faulty reasoning. Detecting and classifying them is a crucial NLP task to prevent misinformation, manipulative claims, and biased decisions. However, existing fallacy classifiers are limited by the requirement for sufficient labeled data for training, which hinders their out-of-distribution (OOD) generalization abilities. In this paper, we focus on leveraging Large Language Models (LLMs) for zero-shot fallacy classification. To elicit fallacy-related knowledge and reasoning abilities of LLMs, we propose diverse single-round and multi-round prompting schemes, applying different task-specific instructions such as extraction, summarization, and Chain-of-Thought reasoning. With comprehensive experiments on benchmark datasets, we suggest that LLMs could be potential zero-shot fallacy classifiers. In general, LLMs under single-round prompting schemes have achieved acceptable zero-shot performances compared to the best full-shot baselines and can outperform them in all OOD inference scenarios and some open-domain tasks. Our novel multi-round prompting schemes can effectively bring about more improvements, especially for small LLMs. Our analysis further underlines the future research on zero-shot fallacy classification. Codes and data are available at: https://***/panFJCharlotte98/Fallacy_Detection. © 2024 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Sharpness-Aware Minimization Activates the Interactive Teaching's Understanding and Optimization 38

Sharpness-Aware Minimization Activates the Interactive Teach...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Xu, Mingwei Cao, Xiaofeng Tsang, Ivor W. School of Artificial Intelligence Jilin University China Singapore College of Computing and Data Science Nanyang Technological University Singapore

Teaching is a potentially effective approach for understanding interactions among multiple intelligences. Previous explorations have convincingly shown that teaching presents additional opportunities for observation and demonstration within the learning model, such as data distillation and selection. However, the underlying optimization principles and convergence of interactive teaching lack theoretical analysis, and in this regard co-teaching serves as a notable prototype. In this paper, we discuss its role as a reduction of the larger loss landscape derived from Sharpness-Aware Minimization (SAM). Then, we classify it as an iterative parameter estimation process using Expectation-Maximization. The convergence of this typical interactive teaching is achieved by continuously optimizing a variational lower bound on the log marginal likelihood. This lower bound represents the expected value of the log posterior distribution of the latent variables under a scaled, factorized variational distribution. To further enhance interactive teaching's performance, we incorporate SAM's strong generalization information into interactive teaching, referred as Sharpness Reduction Interactive Teaching (SRIT). This integration can be viewed as a novel sequential optimization process. Finally, we validate the performance of our approach through multiple experiments. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

AesExpert: Towards Multi-modality Foundation Model for Image Aesthetics Perception 24

AesExpert: Towards Multi-modality Foundation Model for Image...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Huang, Yipo Sheng, Xiangfei Yang, Zhichao Yuan, Quan Duan, Zhichao Chen, Pengfei Li, Leida Lin, Weisi Shi, Guangming School of Artificial Intelligence Xidian University Xi'an China College of Computing and Data Science Nanyang Technological University Singapore

ISBN: (纸本)9798400706868

The highly abstract nature of image aesthetics perception (IAP) poses a significant challenge for current multimodal large language models (MLLMs). The lack of human-annotated multi-modality aesthetic data further exacerbates this dilemma, resulting in MLLMs falling short of aesthetics perception capabilities. To address the above challenge, we first introduce a comprehensively annotated Aesthetic Multi-Modality Instruction Tuning (AesMMIT) dataset, which serves as the footstone for building multi-modality aesthetics foundation models. Specifically, to align MLLMs with human aesthetics perception, we construct a corpus-rich aesthetic critique database with 21,904 diverse-sourced images and 88K human natural language feedbacks, which are collected via progressive questions, ranging from coarse-grained aesthetic grades to fine-grained aesthetic descriptions. To ensure that MLLMs can handle diverse queries, we further prompt GPT to refine the aesthetic critiques and assemble the large-scale aesthetic instruction tuning dataset, i.e. AesMMIT, which consists of 409K multi-typed instructions to activate stronger aesthetic capabilities. Based on the AesMMIT database, we fine-tune the open-sourced general foundation models, achieving multi-modality Aesthetic Expert models, dubbed AesExpert. Extensive experiments demonstrate that the proposed AesExpert models deliver significantly better aesthetic perception performances than the state-of-the-art MLLMs, including the most advanced GPT-4V and Gemini-Pro-Vision. Project Page: https://***/aes-expert/. © 2024 ACM.

关键词： aesthetic critique

来源：评论

学校读者我要写书评

暂无评论

Flow Snapshot Neurons in Action: Deep Neural Networks Generalize to Biological Motion Perception 38

Flow Snapshot Neurons in Action: Deep Neural Networks Genera...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Han, Shuangpeng Wang, Ziyu Zhang, Mengmi College of Computing and Data Science Nanyang Technological University Singapore Singapore Show Lab National University of Singapore Singapore

Biological motion perception (BMP) refers to humans' ability to perceive and recognize the actions of living beings solely from their motion patterns, sometimes as minimal as those depicted on point-light displays. While humans excel at these tasks without any prior training, current AI models struggle with poor generalization performance. To close this research gap, we propose the Motion Perceiver (MP). MP solely relies on patch-level optical flows from video clips as inputs. During training, it learns prototypical flow snapshots through a competitive binding mechanism and integrates invariant motion representations to predict action labels for the given video. During inference, we evaluate the generalization ability of all AI models and humans on 62,656 video stimuli spanning 24 BMP conditions using point-light displays in neuroscience. Remarkably, MP outperforms all existing AI models with a maximum improvement of 29% in top-1 action recognition accuracy on these conditions. Moreover, we benchmark all AI models in point-light displays of two standard video datasets in computer vision. MP also demonstrates superior performance in these cases. More interestingly, via psychophysics experiments, we found that MP recognizes biological movements in a way that aligns with human behaviors. Our data and code are available at link. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SACNN-IDS: A self-attention convolutional neural network for intrusion detection in industrial internet of things

引用

CAAI Transactions on Intelligence Technology 2024年第6期9卷 1398-1411页

作者： Mimonah Al Qathrady Safi Ullah Mohammed S.Alshehri Jawad Ahmad Sultan Almakdi Samar M.Alqhtani Muazzam A.Khan Baraq Ghaleb Department of Information Systems College of Computer Science and Information SystemsNajran UniversityNajranSaudi Arabia Department of Computer Science Quaid-i-Azam UniversityIslamabadPakistan Department of Computer Science College of Computer Science and Information SystemsNajran UniversityNajranSaudi Arabia School of Computing Engineering and the Built EnvironmentEdinburgh Napier UniversityEdinburghUK ICESCO Chair Big Data Analytics and Edge Computing Quaid-i-Azam UniversityIslamabadPakistan

Industrial Internet of Things(IIoT)is a pervasive network of interlinked smart devices that provide a variety of intelligent computing services in industrial *** IIoT nodes operate confidential data(such as medical,transportation,military,etc.)which are reachable targets for hostile intruders due to their openness and varied *** Detection Systems(IDS)based on Machine Learning(ML)and Deep Learning(DL)techniques have got significant ***,existing ML and DL-based IDS still face a number of obstacles that must be *** instance,the existing DL approaches necessitate a substantial quantity of data for effective performance,which is not feasible to run on low-power and low-memory *** and fewer data potentially lead to low performance on existing *** paper proposes a self-attention convolutional neural network(SACNN)architecture for the detection of malicious activity in IIoT networks and an appropriate feature extraction method to extract the most significant *** proposed architecture has a self-attention layer to calculate the input attention and convolutional neural network(CNN)layers to process the assigned attention features for *** performance evaluation of the proposed SACNN architecture has been done with the Edge-IIoTset and X-IIoTID *** datasets encompassed the behaviours of contemporary IIoT communication protocols,the operations of state-of-the-art devices,various attack types,and diverse attack scenarios.

关键词： convolutional neural network deep learning industrial internet of things intrusion detection self-attention

来源：评论

学校读者我要写书评

暂无评论

Effortless learning: leveraging summarization and tabulation for automated note making

引用

Multimedia Tools and Applications 2024年 1-26页

作者： Arivarasi, A. Srinivasan, R. Rajeswari, D. Govindasamy, Alagiri School of Electronics Engineering Vellore Institute of Technology Tamil Nadu Chennai600127 India Department of Computing Technologies School of Computing College of Engineering and Technology SRM Institute of Science and Technology Kattankulathur603203 India Department of Data Science and Business Systems School of Computing College of Engineering and Technology SRM Institute of Science and Technology Kattankulathur603203 India Future Connect Technologies Pvt. Ltd Tamil Nadu Kanchipuram600034 India

In this fast-paced Digital Age, Natural Language Processing (NLP) can prove beneficial in consuming quality information efficiently. With the ever-growing number of learning resources, it is becoming onerous for students to choose the best-suited learning resource. Furthermore, many students find it difficult to understand and retain highly factual information. To overcome these issues, this paper proposes a novel Automated TExt Summarization of scholastic text in Table format (Auto-TEST) to provide information at a glance by mimicking the student’s style of note-making. The Auto-TEST framework is achieved through two steps which is Summarization and Tabulation. Initially, the Extractive Single Document Text Summarization (ESDTS) method is adopted to identify the salient information from the non-fictional (scholastic) text which is extracted and grouped together to form a concise summary. A Named Entity Recognition (NER) method is implemented to extract noun chunks and then arrange them in a tabulation format that aids in memorizing essential facts and figures by enabling effortless retrieval. The performance of the Auto-TEST framework is gauged by using scholastic domain-specific dataset using ROUGE as an evaluation metric. The extracted entities gives an average Recall of 71.40%, an average Precision score of 75.80% and an average F-score of 73.54% for the dataset respectively. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：